466 PROFESSOR KNOTT ON THE STRAINS PRODUCED IN IRON, 



27 : 20. The hypothesis suggested above requires this ratio to be unity. In 

 making this comparison, however, we should bear in mind that the B tube is, 

 because of its shorter length, more uniformly magnetized throughout, and has 

 therefore less leakage of lines of force from the sides than the A tube. This would 

 tend to accelerate the shifting of the maximum to lower fields, an effect which in 

 less uniform fields would be produced by widening the bore of the tube. 



The cross-sections of A V. and B VII. are as 3*24 to *82 ; hence their 

 demagnetizing factors arc as 81 to 82. These tubes should, according to the 

 theory, have their maximum elongations in the same field. As a matter of fact, 

 the field corresponding to the maximum elongation of A V. is somewhat higher 

 than that of B VII. 



A similar comparison of Tubes VII'. and A III., in both of which the field of 

 maximum elongation is 140, brings out 39 : 32 as the ratio of the demagnetizing 

 factors. The deviation of this ratio from unity may also be partly accounted for 

 by the greater uniformity of magnetization in the somewhat shorter A tube. 



The steadiness with which in all cases the field of maximum elongation diminishes 

 as the bore is widened is particularly noteworthy. 



Contrary to the usual experience, I have had no difficulty in measuring the changes 

 of volume of the magnetized material. A glance at the SV columns of Table VI. 

 shows that these changes in the case of iron are by no means insignificant. The 

 corresponding dilatations (8 in Table VII.) have fairly similar values in the A tubes 

 until A VII. is reached, but vary considerably from tube to tube in the B series. 



The tendency is for the cubical dilatation to diminish as the walls of the tube 

 become very thin. Also, both in A VII. and B VII. there is a maximum dilatation 

 in a moderate field. The existence of this maximum need in no way surprise us, for 

 a like peculiarity appears in the longitudinal dilatation. It is rather matter for 

 surprise that there should have been no hint at a maximum cubical dilatation with the 

 other tubes. I have already suggested that the cubical dilatation may be appreciably 

 affected by a vesicular condition in the material, and it is well known that such 

 molecular changes are influenced by the form of the body. Now the thinner the wall, 

 the less chance is there of the presence of vesicular cavities ; and it is also conceivable 

 that in very thin walled tubes there is increased uniformity of magnetization with 

 possibly less leakage of the lines of induction from the sides. Either or both of these 

 considerations seem to give a sufficient explanation of the phenomenon just described. 



§ 7. Curious Behaviour of Iron Tubes under Certain Conditions. — In all 

 experiments in which the iron tubes were enclosed in the brass tube — that is, in 

 experiments of type (3) and (4) of § 2 — a curious effect was observed, of which, so far, 

 I have been unable to find a satisfactory explanation. I can best indicate its nature 

 by transcribing from the experimental note-book the unreduced numbers which 

 measured the volume changes. These represent the number of divisions on the 

 micrometer scale through which the water meniscus in the capillary tube appeared 



